Benefits from Using an Artificial Neural Network as a Prediction Model for Bio-hydrogen Production

The performance of the glucose-based production of H-2 in a batch reactor was predicted by an artificial neural network (ANN). The potential of utilizing an ANN modeling approach to simulate and predict the hydrogen production of Clostridium saccharoperbutylacetonicum N1-4 (ATCC 13564) was investigated. Sixty experimental data records have been utilized to develop the ANN model. In this paper, a unique architecture has been introduced to mimic the inter-relationship between three input parameters: initial substrate concentration, initial medium pH and temperature (10 g/l, 6.0 +/- 0.2, 37 degrees C, respectively). A comparative analysis with a traditional Box-Wilson Design (BWD) statistical model proved that the ANN model output significantly outperformed the BWD model at similar experimental conditions. The results showed that the ANN model provides a higher level of accuracy for the H-2 prediction and fewer errors and that it overcomes the limitation of the BWD approach with respect to the number of records, which merely considers a limited length of stochastic patterns for H-2 prediction.